Frontmatter
| id | 8347 |
| title | Refactor Memory & GitHub Services to Config-based Opt-Out Auto-Sync |
| state | Closed |
| labels | enhancementai |
| assignees | tobiu |
| createdAt | Jan 6, 2026, 11:34 AM |
| updatedAt | Jan 6, 2026, 11:37 AM |
| githubUrl | https://github.com/neomjs/neo/issues/8347 |
| author | tobiu |
| commentsCount | 1 |
| parentIssue | null |
| subIssues | [] |
| subIssuesCompleted | 0 |
| subIssuesTotal | 0 |
| blockedBy | [] |
| blocking | [] |
| closedAt | Jan 6, 2026, 11:37 AM |
Refactor Memory & GitHub Services to Config-based Opt-Out Auto-Sync

Input from Gemini 3 Pro:
✦ I have successfully implemented the config-based "Opt-Out" pattern for Memory Core and GitHub Workflow services.
Changes Implemented:
Memory Core:
- Config: Added
autoSummarize: truetoai/mcp/server/memory-core/config.mjs.- Service: Updated
ai/mcp/server/memory-core/services/SessionService.mjsto checkaiConfig.data.autoSummarizebefore running startup summarization.- SDK: Updated
ai/services.mjsto setMemory_Config.data.autoSummarize = false.GitHub Workflow:
- SDK: Updated
ai/services.mjsto setGH_Config.data.syncOnStartup = false(leveraging the existing config property).Verification: Ran
node ai/examples/debug_session_state.mjs(which uses the SDK). Confirmed that the[Startup] Checking for unsummarized sessions...log message was absent, proving the auto-summarization was successfully disabled. The script was still able to manually trigger summarization using the service methods, confirming that the functionality remains available when explicitly invoked.This ensures a consistent behavior across all AI infrastructure services: "Batteries Included" for MCP Servers (auto-sync/summarize), but "Opt-In" for SDK scripts (no side effects).
Apply the config-based "Opt-Out" pattern (from #8346) to
Memory CoreandGitHub Workflowservices. This ensures that scripts using the AI SDK (ai/services.mjs) do not trigger unintended side effects (auto-summarization or full GitHub syncs) on startup.Plan:
Memory Core:
ai/mcp/server/memory-core/config.mjs: AddautoSummarize: true.ai/mcp/server/memory-core/services/SessionService.mjs: CheckaiConfig.data.autoSummarizeininitAsync.ai/services.mjs: SetMemory_Config.data.autoSummarize = false.GitHub Workflow:
ai/services.mjs: SetGH_Config.data.syncOnStartup = false. (Note:syncOnStartupalready exists inGH_Config).This unifies the behavior across all three AI infrastructure services (KB, Memory, GH).